SFC: Achieve Accurate Fast Convolution under Low-precision Arithmetic

TL;DR

This paper introduces SFC, a novel algebra transform for fast, low-precision convolution that reduces multiplications and maintains accuracy, outperforming existing methods in efficiency and precision.

Contribution

The paper proposes SFC, an innovative algebra transform extending DFT with symbolic computing, enabling efficient, low-precision convolution with correction terms and numerical error analysis.

Findings

Achieves 3.68x reduction in multiplications for 3x3 convolution.

Outperforms Winograd algorithm with similar low numerical errors.

Demonstrates improved efficiency on benchmarks and FPGA while maintaining accuracy.

Abstract

Fast convolution algorithms, including Winograd and FFT, can efficiently accelerate convolution operations in deep models. However, these algorithms depend on high-precision arithmetic to maintain inference accuracy, which conflicts with the model quantization. To resolve this conflict and further improve the efficiency of quantized convolution, we proposes SFC, a new algebra transform for fast convolution by extending the Discrete Fourier Transform (DFT) with symbolic computing, in which only additions are required to perform the transformation at specific transform points, avoiding the calculation of irrational number and reducing the requirement for precision. Additionally, we enhance convolution efficiency by introducing correction terms to convert invalid circular convolution outputs of the Fourier method into effective ones. The numerical error analysis is presented for the first time in this type of work and proves that our algorithms can provide a 3.68x multiplication reduction for 3x3 convolution, while the Winograd algorithm only achieves a 2.25x reduction with similarly low numerical errors. Experiments carried out on benchmarks and FPGA show that our new algorithms can further improve the computation efficiency of quantized models while maintaining accuracy, surpassing both the quantization-alone method and existing works on fast convolution quantization.